How a coding AI works: agents, tokens, tools
Habr has published a useful breakdown of how AI agents for software development actually work. Behind the "magic" is ordinary engineering: an LLM, a system prompt, chat context, tool calls and a loop of repeated requests. The piece shows well why an agent's quality depends not only on the model, but also on the length of the history, token caching, the set of tools and the reasoning mode.
AI-processed from Habr AI; edited by Hamidun News
Habr has published a detailed breakdown of how AI coding agents are built. The text strips away the aura of "magic" and shows that behind the convenient interface sit very concrete mechanics: tokens, a system prompt, tools, dialogue memory and a loop of repeated model calls.
What is an AI agent for programming?
A development agent is a wrapper around a large language model. Inside such a system there is the LLM itself, a hidden system prompt with behavior rules, a list of available tools, and code that runs all of this in a "request → function call → result" loop.
- Nov 1, 2026Local models lose to Opus and Codex: 0.21 on a MacBook versus 1.00 for Codex
- Nov 1, 2026Kimi scores 96.2% on a 1C database analytics test and emerges as a fallback if Claude access is blocked
- Nov 1, 2026OpenAI Rewrites GPT Behavior Rules: Five Values and a Chain of Command
- Oct 31, 2026Author puts his script before an AI jury: the verdict cost 850 thousand tokens
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.